With the growth of network technology, collection of information (data) in a sensor, a camera, a database, and the like connected to a network, and visualization, analysis, and control of the collected information (data) have been widely pursued. On a factory line, for example, for the purpose of real time monitoring, improved production efficiency and quality control, it is necessary to analyze not only structured data such as sensor information and semi-structured data such as device logs but also unstructured data such as an image and a video.
Since the unstructured data such as an image and a video is larger than the structured data in a data volume per entry, it is necessary to adopt transmission/storage schemes for the unstructured data different from those for the structured data. For example, as for the structured data, which is several kilobytes in size per entry, a system that facilitates retrieving the structural data can be constructed by transmitting the structured data as one message on message-oriented middleware, storing the structured data in a relational DB as one record in one entry, and adding an appropriate index to the structured data. However, the unstructured data such as a video is several megabytes to several gigabytes per entry; thus, a method of transmitting the unstructured data by file transfer and storing the unstructured data in a file system, or the like is employed. Alternatively, the video is distributed in a stream type and a concept of “one entry” is not often present at all for the unstructured data; thus, the need arises to determine whether to display the video as it is or to convert the video into files at an appropriate magnitude on a receiving side. It has been necessary to construct a system handling a combination of the unstructured data and the structural data in compliance with requirements of each site on an as-needed basis.
Since the unstructured data and the structured data exhibit different characteristics from each other, fixing data to be collected per application such as a data analysis application and constructing a system dedicated to the application have been widely pursued.
These systems enable a user to easily perform data analysis by associating metadata with the collected data. For example, associating information about an object to be photographed, that is, metadata with video data obtained by photographing the object by a camera makes it possible to display the metadata about the object together with the video and to present the video easier to understand for a viewer.
Examples of such systems include techniques described in JP-2004-185456-A (Patent Document 1) and JP-2008-252410-A (Patent Document 2).
Patent Document 1 describes the following technique. “A communication system for receiving, interpreting, and converting metadata is disposed between communication terminals. Contents can be quickly provided by changing only the metadata without changing the very contents. In addition, an optimum content is selected by referring to various pieces of information such as viewer attribute information, client type information, and advertisement attribute information. The communication system is easily operated by providing a mechanism for selecting, registering, and changing an advertisement insertion policy by an advertiser or an advertisement distribution business operator, a mechanism for automatically determining an advertisement insertion point into the contents, and a mechanism for registering an advertisement attribute.” In this way, Patent Document 1 discloses a method of selecting the optimum content for a receiving side by disposing the communication system that converts meta-information between a transmitting side of transmitting the contents containing metadata and the receiving side.
Furthermore, Patent Document 2 describes the following technique. “When a user designates an object on a screen by an object designation section 26 of a content display apparatus 14, identification information about the object and status data about a video content at this time are transmitted to an information providing server 12. A candidate selector 32 of the information providing server 12 performs retrieval in a reference frequency storage section 40 on the basis of the status data, and selects metadata to be displayed in accordance with points indicative of a weight of a display frequency preset in response to a status and display restrictions. A display information transmitter 36 reads the metadata to be displayed from a metadata storage section 18, and transmits the metadata to the content display apparatus 14.” In this way, Patent Document 2 discloses a method of selecting appropriate metadata from the information providing server in response to identification information about the object when the object is designated in the screen through which a viewer is viewing the video, and transmitting the metadata to the content display apparatus.